Use of an ambient ionization flowing atmospheric-pressure afterglow source for elemental analysis through hydride generation

Abstract

An ambient mass spectrometry ionization source based on an atmospheric-pressure flowing afterglow has been coupled to a Mattauch-Herzog mass spectrograph capable of simultaneous acquisition of a range of mass-to-charge values by means of a Faraday-strip array detector. The flowing afterglow was used as the ionization pathway for species produced by hydride generation. This ionization strategy circumvents problems, such as discharge instabilities or memory effects, induced by introducing the gaseous sample into the discharge. The generated spectra show both atomic and molecular peaks; calibration curves were calculated for both peak types with limits of detection for arsenic below 10 ppb. This study demonstrates the ability to use an ambient mass spectrometry source, commonly used for molecular analyses, for the detection of gas phase elemental species with the possibilty of performing speciation by coupling with a separation technique.

---

Introduction

Ambient mass spectrometry (AMS) has become a very active area of research through the introduction of desorption electrospray ionization (DESI)¹ and direct analysis in real time (DART).² These techniques allow the direct desorption/ionization of a very broad range of samples, and with no sample preparation.^1,3–11 By means of predominantly proton or charge transfer ionization mechanisms, these sources produce simple mass spectra in which the molecular or protonated molecular ion generally dominates.³ Overall, AMS sources are attractive because of high sample throughput, ease of use, low initial costs, and low operating expense.

Recently, a new ambient ionization source has been developed that utilizes the flowing afterglow of a helium atmospheric-pressure glow discharge (GD) to achieve desorption and/or ionization.^12,13 Named the Flowing Atmospheric-Pressure Afterglow (FAPA), this source uses a beam of highly energetic helium metastable (He^*) species to create a high flux of reagent ions, such as protonated water clusters, nitrogen dimer ions, and oxygen ions, to efficiently ionize analyte species that are placed within the afterglow region. This ionization mechanism, comparable to chemical ionization, results in simple mass spectra with little or no fragmentation, just as with other ambient ionization sources. Generally, the protonated molecular ion is formed, but spectra are often difficult to interpret due to the formation of clusters and adducts. As a result, ambient ionization sources generally are not used for elemental analysis; however, with a suitable sample-introduction method atomic information can be acquired.

Hydride generation (HG) enables the production of gaseous hydride species and is a standard method¹⁴ for several elements, including arsenic, germanium, antimony, selenium, tellurium, and others.¹⁵ These volatile hydrides are produced by mixing an analyte-containing acid solution, usually hydrochloric acid, and a solution of sodium borohydride. Upon mixing these components, the reaction in equation 1 proceeds. Although hydrogen is produced, there is ongoing debate about its form.¹⁶ The hydrogen can then combine with an anaylte species, such as those listed above, to form a volitile hydride, as in equation 2, where A is the analyte species, m isthe valence of the analyte A in solution, and n is the valence of analyte A in the hydride.¹⁷

BH₄⁻ + 3H₂O + H⁺ → H₃BO₃ + 8H (1)

A^m+ + (m + n)H → AH_n + mH⁺ (2)

A common example is the use of arsenic as the analyte, in which arsine is the product. Coincidentally, many of the hydride-forming elements are important in speciation analysis due to the toxicity of their various forms to living systems.¹⁸ As a result, HG has been utilized as a sample introduction method for such analyses.¹⁹ In addition, HG provides several advantages: (1) high analyte transport effieciencies, (2) low ionization source power requirements because of the lack of a necessary desolvation step, (3) efficient separation of analytes from matrix species due to the selectivity of the HG process, and (4) the ability to maintain chemical speciation information during the derivitization step in cases such as mercury, germanium, and tin.^15,18

Previously, HG has been coupled to mass spectrometry through the use of inductively coupled plasma (ICP)²⁰ or GD²¹ ionization sources. With these sources, the hydride species must be introduced into the plasma, which can change the plasma characteristics.^21–26 This problem can become even more serious when HG is used in conjunction with transient sample introduction techniques in which the gaseous sample composition continuously changes. An example is chromatographic separations, which are generally essential for speciation analysis.

In the present paper, the use of HG with the FAPA source is used to achieve the advantages of HG while circumventing problems caused by sending high concentrations of hydride species into a plasma. The aim is to evaluate the use of a simple, low cost ionization source as an alternative method for performing elemental analyses by HG. Furthermore, the use of HG with an ambient ionization source provides a step towards coupling such sources to gaseous sample introduction techniques such as gas chromatography.

Experimental

Sample introduction system

The hydride generation apparatus used in these studies has been describe previously,²⁴ so only a brief description and relevant changes will be described here. The HG process was performed by mixing a solution of 5% sodium borohydride, stabilized in 5% sodium hydroxide, with a 1 M hydrochloric acid solution in a glass tee. The HCl solution contained the analyte of interest. The resulting mixture was pumped at a rate of 0.5 or 1.0 mL/min through an approximately 150 cm PTFE reaction coil using a 4-head peristaltic pump (Gilson, Middletown, WI). The coil was used to ensure complete reactant mixing and hydride species formation. The combined gas and solution mixture was then passed through a gas-liquid separator (GLS) filled with glass beads to provide ample surface area. A sweep gas inlet to the GLS was closed off during these experiments. After the GLS, a peristaltic pump (Gilson, Middletown, WI) was used to drive the liquid to waste, while the hydride vapor flowed through a condensing column, held at a temperature of ∼5 °C by means of a recirculating chiller (ThermoNeslab M75, Thermo-Electron Corporation, Waltham, MA), to remove residual water vapor from the gas mixture. The remaining hydride vapor was passed through a 915 µm i.d. glass capillary to the FAPA source. The glass capillary was held in an x,y,z fiber-optic positioner (Newport Corporation, Irvine, CA) in order to tune its position in the afterglow. A diagram of the FAPA and MHMS interface can be found in Fig. 1.

Fig. 1 Diagram of the FAPA system and the MHMS interface.

Flowing atmospheric-pressure afterglow

The FAPA consists of a tungsten-pin anode and brass-plate cathode. The two electrodes are held in a Teflon^®cell that has a side inlet for helium (1–3 L/min, 99.999% ultra high purity, Airgas, Radnor, PA), which served as the discharge gas. The discharge was sustained by a high voltage (EH Series, Glassman High Voltage, Inc., Whitehouse Station, NJ) power supply connected in series with the two electrodes. Three 1.25 kΩ, 50 W resistors were placed in series in the return path to the power supply. In all cases the discharge was sustained in a current-controlled mode. The discharge cell was held in a locally built x,y-translation stage to tune its position with respect to the mass spectrometer interface. The x,y-translation stage was placed on optical rails to allow optimization of the z position (distance from the discharge cell to the MHMS interface).

Mattauch-Herzog mass spectrograph and focal plane camera

The MHMS and FPC have been described in detail previously,^27–30 so only a brief description and pertinent changes for the current work will be given here.

Slight modifications to the MHMS-ICP interface were made in order to efficiently sample ions from the FAPA source. These changes include the use of a traditional mass spectrometer sampling cone in which the conical portion was machined off, referred to here as the sampling plate. A flat front plate was used to replace the cone vacuum orifice, and the front plate was electrically isolated from the sampling plate by means of a Viton^® o-ring. Additionally, an electrically isolated aluminum ion lens was placed between the sampling plate and the skimmer cone. The sampling plate and the skimmer cone are electrically connected and their potential was held at 1 kV by a high-voltage power supply (SL-10, Spellman, Plainville, NY). The potential of the front plate can be varied by using the combination of two power supplies (0–6 kV, SL-30, Spellman, Plainville, NY and 0–60 V dc, 6024A, Hewlett Packard (Agilent), Santa Clara, CA). The low-voltage power supply was used for fine tuning the potential of the front plate. Because of the large potential difference between the FAPA and the MHMS interface, no ions were detected unless the FAPA potential was referenced to the potential of the front plate instead of to ground. Therefore, both the FAPA and the fine tuning front-plate power supplies were referenced to the high voltage front-plate power supply. The ion lens between the sampling plate and the skimmer cone was maintained at the same potential as the front plate high-voltage power supply. No changes after the skimmer cone were made to the MHMS instrument.

The second-generation FPC detector was used in the current studies. This version consists of 128 gold Faraday strips, each of which is 45-µm wide; the strips are set on 50-µm centers. The gain of each Faraday strip can be independently selected between two levels. The high gain level was used for all experiments except where indicated. Furthermore, nondestructive readouts were used to minimize read noise associated with the detector electronics.

Time-of-flight mass spectrometer

Because the MHMS can measure only the low (atomic) mass-to-charge range, a LECO HT Unique^® (LECO Corp. St. Joseph, MI) time-of-flight (TOF) mass spectrometer was used to obtain full mass spectra for the molecules created in the HG process. Only two differences were needed in the source configuration during coupling with the Unique^®spectrometer. First, the sampling cone was held at a relatively low potential (∼60 V), so there was no need to reference source power supplies with the inlet potential. Second, the spectrometer was designed for conventional APCI and ESI sources, so the vacuum system could not cope with high helium flow rates (>0.9 L/min). As a result, the FAPA had to be operated at lower flow rates. The consequences of this change will be addressed in the Results and Discussion section.

Sample preparation

Sodium borohydride solutions were prepared fresh daily and were stabilized with sodium hydroxide. Except where noted, solutions of sodium borohydride were prepared by dissolving 5 g of sodium borohydride (98.5% powder, Sigma-Aldrich, St. Louis, MO) and 5 g of sodium hydroxide (97.0% pellets, EM Science, Gibbstown, NL) in 100 mL of distilled, deionized water. Analyte solutions were prepared in 1 M hydrochloric acid obtained through dilution of concentrated hydrochloric acid (ACS Grade, Mallinckrodt Baker, Inc., Phillipsburg, NJ). Analyte stock solutions were prepared at the 10 ppm level and serial dilutions with deionized water were made to obtain lower concentrations.

Results and discussion

FAPA background characterization

The mass spectral background of the FAPA source was determined in the absence of the HG system and is shown in Fig. 2. The background spectrum is shown in the form of a bar graph due to the nature of the current MHMS-FPC system. At present, the FPC is only wide enough to acquire between 2 and 30 m/z values simultaneously, depending on the mass range of observation. For this reason, it is necessary to vary the magnetic field to acquire an entire spectrum. Therefore, only a narrow range of background peaks was focused on the FPC at a time, and the peaks were individually integrated. The height of the bars in the graph represents the integrated peak area at each m/z value. The major background species from the FAPA are protonated water clusters. These cluster species are believed to ionize analytes in the afterglow region through proton transfer reactions,¹³ analogous to atmospheric-pressure chemical ionization (APCI) mechanisms.^31,32 In addition to the water clusters, other ions from atmospheric components such as NO⁺ and O₂⁺ are also present in the background. Furthermore, very minor peaks are observable at nearly all m/z values below m/z 200. These peaks are believed to be clusters or adducts of various atmospheric components or contaminants in the gas. No attempt was made to further characterize these peaks. Future experiments will be aimed at controlling the atmosphere in order to reduce the effects of such peaks. For all following experiments, data were blank subtracted using the HG system with blank solutions of 1M HCl and 5% NaBH₄.

Fig. 2 Background spectrum of the FAPA source without the HG system. The major peaks are protonated water clusters and ionized atmospheric components.

Optimization of hydride generation

No attempt was made to fully optimize the HG introduction system for the present study. Instead, previous experience with the system was used as a guide. However, two parameters that were experimentally optimized were acid concentration and sample flow rate. The ⁷⁵As⁺, ⁷⁵AsH⁺, and ⁷⁵AsH₂⁺ signals were monitored for both 1 M and 2 M HCl acid concentrations; a plot of these data is shown in Fig. 3. The signal for these three species drops by approximately a factor of two when the 2 M HCl solution is used. It was hypothesized that the higher acid concentration produced too much H₂ for the FAPA to accept; however, this hypothesis was not verified. In addition, the sample flow rate was varied from 0.5 to 1 mL/min; no variation in the signal levels for ⁷⁵As⁺, ⁷⁵AsH⁺, and ⁷⁵AsH₂⁺ was observed. Therefore, the higher flow rate was used for all subsequent experiments in order to minimize the washout time of the HG system.

Fig. 3 Spectrum containing the As⁺, AsH⁺, amd AsH₂⁺ peaks obtained using two different acid concentrations for HG. The higher acid concentration yields a reduction in signal level.

Hydride generation of arsenic

As spectrum. Arsenic was the analyte of choice for the full characterization of the experimental system in part because of its importance in many speciation studies. First, a mass spectrum was acquired to characterize the peaks that are observed upon introduction of arsine into the FAPA. For this a standard solution of 10 ppm As was used. The mass spectum (m/z 90 to 260) is shown in Fig. 4a. This mass spectrum was produced by splicing together several smaller mass windows. The strongest As peak was found to be AsO⁺; however, As⁺ was also observed as shown in Fig. 4b. In addition to these As peaks, numerous As cluster species were observed, many of which correspond closely to the spectrum obtained by Mulligan, et. al., who used APCI to determine arsine.³³ Assignments shown in Fig. 4 are based on those made in reference 33; no experimental attempt was made to verify these cluster assignments.

Fig. 4 Spectra of the clusters formed during the HG of arsenic by using the FAPA-MHMS-FPC system. a) Full spectrum accumulated from many mass windows by varying the magnetic field of the MHMS, and b) zoomed-in spectrum acquired using only two mass windows. Each point in b corresponds to a different Faraday strip of the FPC array detector.

Precision and washout time. A solution of 10 ppm As was sent through the HG system for a period of 30 min while the signal level for m/z 91 (AsO⁺) was monitored at 1 min intervals. The precision of these 30 measurements was found to be 7.4% RSD, most likely limited by instabilities, such as bubble formation, in the HG process. After the 30-min time period, the 10 ppm As solution was replaced with a 1M HCl blank solution, while continuing to monitor m/z 91. The signal dropped to <10% of the steady-state value in 5 minutes and to the baseline value after about 20 min. The 10 ppm As solution was again introduced after the washout; the signal returned to the steady state value in about 1 min. These data are displayed in Fig. 5. The washout time of the system is somewhat longer than desirable because of the large surface area created by the glass beads in the gas/liquid separator. Presumably, faster washout times could be achieved by reducing the surface area in the separator or using faster flow rates during the washout period.

Fig. 5 Washout curve and precision for the HG-FAPA System. Complete washout occurs in about 20 minutes. The precision of the first 20 points is 7.4% RSD.

Noise characterization. System noise was characterized from the noise power spectra (NPS) for the As⁺ analyte peak, the background (H₂O)₃H⁺ peak and the detector, the latter under dark current (no ionization source) conditions. The noise spectra were determined by collecting signal data at 500 µs intervals over a 10 s total integration period. The low-gain setting of the FPC was used to avoid saturation of the Faraday strips; however, it was experimentally verified that the noise of the detector was not significantly different between the low- and high-gain settings. The NPS, shown in Fig. 6, indicate that the dominant noise features are the harmonics of the United States power line frequencies (120 Hz, 240 Hz, etc.). The largest of these peaks appear in the As⁺ analyte noise power spectrum, indicating the presence of some additional line noise in the sample introduction system. However, the intensities of these peaks are nearly the same between the source background and dark current NPS, indicating that no additional line noise was introduced by the FAPA source. In addition to the line noise, some flicker (1/f) noise is present; however, it is nearly identical among the three spectra. Johnson noise from the FPC has been minimized through the use of a Peltier cooler, as previously described.³⁴

Fig. 6 Noise power spectra for the As⁺ peak (m/z 75), a background peak (m/z 55), and the dark current from the detector. The major features are from power line noise and the largest noise level occurs for the analyte peak. Some flicker noise is present and is also greatest for the analyte peak.

Calibration curves. Calibration curves, shown normalized in Fig. 7, were determined for the As⁺, AsO⁺, and As₂O₃⁺analyte peaks. As analyte concentration goes up, the As⁺ and AsO⁺ calibration curves become very nonlinear and begin to roll off. When this nonlinearity begins, a signal for As₂O₃⁺ appears and increases with greater analyte concentrations. This observation suggests that higher analyte concentrations promote the formation of clusters within the afterglow or first vacuum stage regions.

Fig. 7 Calibration curves for three analyte peaks of arsenic. The As⁺ and AsO⁺ curves roll off at higher concentrations, whereas the As₂O₃⁺ shows upward curvature. This behavior is due to increased formation of clusters at higher concentrations. The plots have been normalized, so slopes should not be compared quantitatively.

The linear portion of the calibration plots, at lower concentrations, were used to calculate detection limits (3σ), which were found to be 7 ppb, 2 ppb, and 30 ppb for As⁺, AsO⁺, and As₂O₃⁺, respectively. These values are slightly worse than or on par with a gas sampling GD that was used in our laboratory with the HG system;²¹ yet, they are about 100 times worse than values obtained with an atmospheric-pressure glow discharge that was directly coupled to the interface of a TOFMS.²⁴ However, in that study the hydride species were sent through the discharge rather than into the afterglow of the discharge. When the analytes are introduced into the discharge, only atomic As was detected, so only one peak had to be analyzed. In the present study, the As species are spread across several cluster peaks, making quantification more difficult. Certainly, work in the future should focus on reducing these cluster species if elemental analysis continues to be a goal. When the detection limits here are compared to those from a more commonly used method, such as HG-ICPMS, the values are between 10–10000 times worse.^21,35 However, considering the added complexity and cost of operating an ICP, this technique might be desirable if sensitivity is not the dominant concern.

Hydride generation of germanium and antimony. Hydride generation of germanium and antimony was also investigated; the mass spectra obtained with the MHMS-FPC instrument are shown for each element in Fig. 8a and 8b, respectively. For these elements, only the spectra for the atomic species were collected. Assignments shown in Fig. 8 are based purely on mass, and no experimental attempt was made to identify the specific species or the ratio of species at a given mass. One complication with the described method is the formation of adducts and clusters. This process not only divides the analyte signal among many m/z values but also complicates any isotope data that can be abstracted.

Fig. 8 Spectra from the HG of (a) Ge and (b) Sb collected using the MHMS-FPC. Two mass windows were used to collect the Ge isotopes and pieced together where indicated. Both spectra show atomic species and hydrides.

As, Ge, and Sb spectra characterization with FAPA-TOFMS. A LECO Unique^® TOFMS was used to generate full mass spectra in order to view simultaneously all species formed in the hydride generation and the ionization processes. Mass spectra for As, Ge, and Sb are shown in Fig. 9a, 9b, and 9c, respectively. For the As spectrum, many of the same clusters are observed with the TOFMS as with the MHMS-FPC. However, some different clusters appear, which is likely due to the different operating conditions of the FAPA source. At lower helium flow rates, dimerization signals strongly increased (data not shown). This elevated appearance is likely due to a drop in temperature and energy of the afterglow, which has been demonstrated experimentally.³⁶ Another factor that could play a significant role in causing the dissimilarities between the spectra observed with the two types of mass analyzers are different interface geometries. The LECO Unique^®spectrometer uses a capillary interface, whereas the MHMS uses sampling orifices. Furthermore, different optimal lens potentials and/or variations in the atmosphere could also contribute. In the Ge and Sb spectra, the clusters are due mainly to the addition of water molecules or oxygen atoms. In the Ge and Sb TOFMS spectra, it is possible to compare the theoretical and experimental isotopic patterns using the GeOH⁺ and SbO⁺ peaks, since they reside in relatively clean areas of the spectra and no proton additions were observed. The evaluation showed that the isotopic distributions are generally only 1–5% different from the expected values.

Fig. 9 Spectra from the HG of (a) As, (b) Ge, and (c) Sb taken with the FAPA and the Unique^® TOFMS. Many of the same clusters are observed in the two mass analyzers for As (c.f Figs. 4 & 8). The clusters observed for Ge and Sb are mainly from addition of water molecules or oxygen atoms.

Conclusions

A hydride generation system has been coupled to an atmospheric-pressure GD ambient ionization source. This union demonstrates not only the ability to couple a gaseous sample introduction system to ambient mass spectrometry, but to obtain elemental information as well. The HG system provides high selectivity and could be coupled with liquid chromatography to perform speciation analysis for analytes such as arsenic. In addition, the system can provide quantitative information; however, care must be taken when considering the linear working range due to the formation of clusters at higher concentrations. This successful coupling would suggest that other gaseous sample introduction techniques, such as gas chromatography, could be effectively coupled with such an ionization source. Furthermore, with the soft ionization of the FAPA, future studies could be aimed at obtaining chemical species information by regulating the atmosphere in which the source is operating, and thereby controlling the degree of adduct or cluster formation. Controlling the atmosphere and the cluster formation would also be desirable in optimizing the sensitivity of the described system.

Acknowledgements

Support for this work was provided by the U.S. Department of Energy, Office of Nonproliferation Research and Engineering. Pacific Northwest National Laboratory is operated by Battelle Memorial Institute for the Department of Energy under Contract DE-AC06-76RLO-1830. José A. C. Broekaert would like to thank the DFG for travel funds. The authors would also like to thank LECO Corporation for the loan of the Unique(R) TOFMS instrument.

References

1. Z. Takats, J. M. Wiseman, B. Gologan and R. G. Cooks, Science, 2004, 306, 471–473.
2. R. B. Cody, J. A. Laramee and H. D. Durst, Anal. Chem., 2005, 77, 2297–2302.
3. R. G. Cooks, Z. Ouyang, Z. Takats and J. M. Wiseman, Science, 2006, 311, 1566–1570.
4. F. M. Fernandez, R. B. Cody, M. D. Green, C. Y. Hampton, R. McGready, S. Sengaloundeth, N. J. White and P. N. Newton, ChemMedChem, 2006, 1, 702–705.
5. K. Kpegba, T. Spadaro, R. B. Cody, N. Nesnas and J. A. Olson, Anal. Chem., 2007, 79, 5479–5483.
6. C. Petucci, J. Diffendal, D. Kaufman, B. Mekonnen, G. Terefenko and B. Musselman, Anal. Chem., 2007, 79, 5064–5070.
7. C. Y. Pierce, J. R. Barr, R. B. Cody, R. F. Massung, A. R. Woolfitt, H. Moura, H. A. Thompson and F. M. Fernandez, Chem. Commun., 2007, 807–809.
8. M. S. Bereman, T. I. Williams and D. C. Muddiman, Anal. Chem., 2007, 79, 8812–8815.
9. I. Cotte-Rodriguez, C. C. Mulligan and G. Cooks, Anal. Chem., 2007, 79, 7069–7077.
10. G. Huang, H. Chen, X. Zhang, R. G. Cooks and Z. Ouyang, Anal. Chem., 2007, 79, 8327–8332.
11. C. C. Mulligan, D. K. MacMillan, R. J. Noll and R. G. Cooks, Rapid Commun. Mass Spectrom., 2007, 21, 3729–3736.
12. F. J. Andrade, J. T. Shelley, W. C. Wetzel, M. R. Webb, G. Gamez, S. J. Ray and G. M. Hieftje, Anal. Chem., 2008, 80, 2654–2663.
13. F. J. Andrade, J. T. Shelley, W. C. Wetzel, M. R. Webb, G. Gamez, S. J. Ray and G. M. Hieftje, Anal. Chem., 2008, 80, 2646–2653.
14. A. E. Greenberg, R. R. Trussell, L. S. Clesceriand M. A. H. Franson, eds., Standard Methods for the Examination of Water and Wastewater, American Public Health Association, Washington, D.C., 1985.
15. P. Pohl, Trends Anal. Chem., 2004, 23, 87–101.
16. A. R. Kumar and P. Riyazuddin, Anal. Sci., 2005, 21, 1401–1410.
17. Hydride Generation Atomic Absorption Spectrometry, ed. J. T. Dedina and L. Dimiter, John Wiley & Sons, Ltd, West Sussex. 1995.
18. R. Cornelis, Handbook of Elemental Speciation: Techniques and Methodology, John Wiley and Sons, Ltd., West Sussex, England, 2003.
19. A. R. Kumar and P. Riyazuddin, Int. J. Environ. Anal. Chem., 2007, 87, 469–500.
20. M. J. Powell, D. W. Boomer and R. J. McVicars, Anal. Chem., 1986, 58, 2864–2867.
21. W. C. Wetzel, J. A. C. Broekaert and G. M. Hieftje, Spectrochim. Acta, Part B, 2002, 57B, 1009–1023.
22. M. Grotti, C. Lagomarsino and J. M. Mermet, J. Anal. At. Spectrom., 2006, 21, 963–969.
23. A. Menendez, J. Pisonero, R. Pereiro, N. Bordel and A. Sanz-Medel, J. Anal. At. Spectrom., 2003, 18, 557–563.
24. W. C. Wetzel, F. J. Andrade, J. A. C. Broekaert and G. M. Hieftje, J. Anal. At. Spectrom., 2006, 21, 750–756.
25. V.-D. Hodoroaba, V. Hoffmann, E. B. M. Steers and K. Wetzig, J. Anal. At. Spectrom., 2000, 15, 1075–1080.
26. V.-D. Hodoroaba, E. B. M. Steers, V. Hoffmann and K. Wetzig, J. Anal. At. Spectrom., 2001, 16, 43–49.
27. J. H. Barnes, G. D. Schilling, R. Sperline, M. B. Denton, E. T. Young, C. J. Barinaga, D. W. Koppenaal and G. M. Hieftje, Anal. Chem., 2004, 76, 2531–2536.
28. J. H. I. V. Barnes, G. D. Schilling, M. B. Denton, D. W. Koppenaal and G. M. Hieftje, J. Anal. At. Spectrom., 2003, 18, 1015–1018.
29. T. W. Burgoyne, G. M. Hieftje and R. A. Hites, J. Am. Soc. Mass Spectrom., 1997, 8, 307–318.
30. A. K. Knight, R. P. Sperline, G. M. Hieftje, E. Young, C. J. Barinaga, D. W. Koppenaal and M. B. Denton, Int. J. Mass Spectrom., 2002, 215, 131–139.
31. J. Sunner, G. Nicol and P. Kebarle, Anal. Chem., 1988, 60, 1300–1307.
32. M. M. Shahin, J. Chem. Phys., 1966, 45, 2600 -&.
33. C. C. Mulligan, D. R. Justes, R. J. Noll, N. L. Sanders, B. C. Laughlin and R. G. Cooks, Analyst, 2006, 131, 556–567.
34. G. D. Schilling, F. J. Andrade, J. H. Barnes, R. P. Sperline, M. B. Denton, C. J. Barinaga, D. W. Koppenaal and G. M. Hieftje, Anal. Chem., 2006, 78, 4319–4325.
35. B. Klaue and J. D. Blum, Anal. Chem., 1999, 71, 1408–1414.
36. J. Shelley, F. Andrade, S. Ray, J. Wiley and G. Hieftje, 34th Annual Conference of the Federation of Analytical Chemistry and Spectroscopy Societies, Memphis, TN, 2007.

---